XM4D: 4D Reconstruction through Certifiable Geometric Optimization
cs.CV, cs.RO, math.OC
Submitted: 2026-11-05
Updated: 2026-11-05
Project page: https://computationalrobotics.seas.harvard.edu/XM4D
Terminology
Sources
- A Riemannian low-rank method for optimization over semidefinite matrices with block-diagonal constraints
- MoBluRF: Motion Deblurring Neural Radiance Fields for Blurry Monocular Video
- SAM 3: Segment Anything with Concepts
- SceneAgent: An Agentic Pipeline from 3D and 4D Captures to Simulation Environments
- RTMPose: Real-Time Multi-Person Pose Estimation based on MMPose
- CoWTracker: Tracking by Warping instead of Correlation
- Depth Anything 3: Recovering the Visual Space from Any Views
- GRASP: Gradient-Guided Sampling for Contact-Rich Physics-Based Motion Retargeting
- HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural Radiance Fields
- D-NeRF: Neural Radiance Fields for Dynamic Scenes
- Embodied Hands: Modeling and Capturing Hands and Bodies Together
- The Replica Dataset: A Digital Replica of Indoor Spaces
- SIM-Sync: From Certifiably Optimal Synchronization over the 3D Similarity Group to Scene Reconstruction with Learned Depth
- Advances in Global Solvers for 3D Vision
- Learning Generalizable Manipulation Policies with Object-Centric 3D Representations
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs